Question

我正在使用sjango，并且尝试使用带有tensorflow后端的Keras预测图像，但是我遇到此错误：

line 3669, in _as_graph_element_locked
    raise ValueError("Tensor %s is not an element of this graph." % obj)
ValueError: Tensor Tensor("dense_6/Softmax:0", shape=(?, 50), dtype=float32) is not an element of this graph.
[28/Apr/2019 16:54:53] "POST /facture/upload/ HTTP/1.1" 500 133945

这是我的代码：

#Loading the model 
pwd = os.path.dirname(__file__)
with open(pwd+'/ModelML/model_architecture19.json', 'r') as f:
    model = model_from_json(f.read())
# Load weights into the new model
model.load_weights(pwd+'/ModelML/model_weights19.h5')



roi = cv2.cvtColor(roi,cv2.COLOR_BGR2GRAY)
            ret2, roi = cv2.threshold(roi, 127, 255, cv2.THRESH_BINARY_INV)    
            roi = cv2.resize(roi, (IMG_SIZE, IMG_SIZE)) # Resize the image
            roi = roi.reshape(1,IMG_SIZE, IMG_SIZE,1)
            #normalize image
            roi = roi /255 


            graph = tf.get_default_graph()
            with graph.as_default():
                pred =model.predict(roi)

错误来自“模型预测”的最后一行

预先感谢您的帮助

Answer 1

graph = tf.get_default_graph()是否在函数或方法内？

在model.load_weights(pwd+'/ModelML/model_weights19.h5')之后将其右移怎么办？

这些行

pwd = os.path.dirname(__file__)
with open(pwd+'/ModelML/model_architecture19.json', 'r') as f:
    model = model_from_json(f.read())
# Load weights into the new model
model.load_weights(pwd+'/ModelML/model_weights19.h5')
graph = tf.get_default_graph()

应该在任何类或函数之外。

预测时存在张量流错误的Keras

1 个答案: